Document not found! Please try again

RA:RE EARTHS. I. VAPORIZATIOS OF Ls20a AND Sd,03

one e.v. Dissociation energies estimated for other rare earth monoxides by Birge-Sponer3 or modified. Birge-Sponer4 extrapolations of the observed vib...
1 downloads 0 Views 683KB Size
1400

H. W.GOLDSTEIN, P. K.WALSHA N D D. WHITE

endothermic heat, of hydrolysis of equation ‘3 is added to parts one and two. Consequently, AE for the solutiori process in carbonate solution will be greater than in hydroxide. The first part of the process is density iridepeadcrlt while the second part is apparently more

Vol. 65

strongly density dependent when canform. Acknowledgments.-‘rhe authors \vi& to thank G. T. Kohman and M. Tannenbaum for advice arid elicouragement in this work. R. A. L. also wishes to acknomledge discussions with E. U. Franck and R. Mosebach.

RA:RE EARTHS. I. VAPORIZATIOS OF Ls20a AND Sd,03: DISSOCISTION ENERGIES OF G-4SEOUS La0 AND NdOl B Y HAROLD IT.GOLDSTEIN, PATRICK 3.JfrALSH

.4SD

DAVID WHITE

Cryogenic Luborutory, Department of Chemistry, ?’he Ohio State University, Columbus, Ohio Recehed March 19. 1981

The vaporization of the rare earth oxides, La208 and Ndz03, a t elevated temperatures has been studied by a combination of Knudsen effuaiion and mass spectrometric techniques. Both vaporize almost stoichiometrically to the monoxide and oxygen. The heizta of formation, AHoa, in kcal. mole-’, and dissociation energies, Doo, in e.v. are: Lao, -29.8 2~ 4,8.08 f 0.2; NdO, -30.0 f 6, 7.18 f 0.3, respectively.

Introduction The question of the identity and stability of the vapor species formed in lanthanon-oxygen systems has been of considerable interest for many years and has becoime increasingly important as rare earth compounds have become more readily available. The existence of gaseous monoxides of several of the rare earth elements has been inferred from stellar and arc spectraza and L a 0 has been observed in the mass spectrometer.2b Only in the latter invwtigation has a dissociation energy been determined with an uncertainty of less than one e.v. Dissociation energies estimated for other rare earth monoxides by Birge-Sponer3 or modified Birge-Sponer4 extrapolations of the observed vibrational levels have such large uncertainties connected with them ( A l - 3 e.v.) that the thermodynamic properties of chemical systems a t elevated temperatures containing the elements of these compounds cannot be determined with sufficient accuracy for most purposes. The positive identification of the monoxides or other simple gaseous oxides of the rare earth metals and the determination of their thermodynamic properties are also of considerable interest from the chemical standpoint. The heats of formation of the solid sesquioxides of these metals are unusually large6; in terms of enthalpy of formation per oxygen atom, they are equaled only by Li20 and some of the Group I1 oxides.6 It is possible that the origin of this strong bonding energy can be inferred from the properties of the gaseous species formed on vaporization of the sesquioxides. The gaseous (1) This work wa3 supported by the Air Force Office of Scientific Research. (2s) B. Rosen, “Donnees Spectroscopiques Concernant les Molecules Diatomiques,” Herman et Cie, Paris, 1951. (b) U‘. A. Chupkn, 11. G . Inghrsm a n d R. F. Porter, J . Chem. Phys., 24, 792 (1956). (3) G. Herzberg, ‘Molecular Spectra a n d Molecular Structure. I Spectra of Diatomic Molecules.” D. Van Nostrand Co., Inc., Princeton, N. J.. 1957. (4) A. G . Gaydon, “Dissociation Energies a n d Spectra of Diatomic hloiecules.” Chapman a n d Hall, Ltd., London, 1953. (5) E. J. Huber, Jr., E. L. Head a n d C. E. Holley, Jr., J . Phys. ChPm., 64, 1708 (19CO) and earlier papers. (C) J. P. (‘oughlin, r.S Bur. Mines Bull., 542 (19.54).

oxides of the rare earths (in particular the monoxides) constitute a unique set of compounds in the sense that the electronic structure of the elements suggests that the chemical bonding should be nearly the same for all. It is, however, well known that there exist large variations in the binding energies4 of these compounds. Before any generalization can be made concerning these variations, it is necessary to establish quantitatively their magnitude In the present paper the dissociation energies of gaseous L a 0 and KdO are determined through Knudsen effusion and mass spectrometric measurements of the dissociation pressures of the corresponding solid sesquioxides. This is the first paper of a series concerned with the determination of the thermodynamic properties of gaseous rare earth oxides. Experimental Materials.-The lanthanum and neodymium oxides were the same materials previously used for heat capacity measurements.’ They were dried before use by firing a t 1000O in air. Apparatus and Procedure.-The effusion experiments were carried out in vacuum induction furnaces of the type described by Hoch and Johnston.* Tantalum and tungsten crucibles, machined bv Fanstcel Metallurgical Company, were used as Knudsen cells. These were right circular cylinders I/*‘‘ 0.d. by ”4’’ high; the tantalum crucibles were S / g “ i.d., the tungsten l/*‘‘. Each was covered with a thick lid of the same material machined to b t inside the crucible (with a tolerance of 0.002”) to a depth of l/Jz”. A 1/168 diameter hole, drilled axially through the lid, served as the effusion orifice. In the calculations, dimensions measured to f 0.001 ” and corrected for thermal erpansiong werc used. Temperatures were measured by Biphting directly into the effusion orifice, through an optical flat a t the top of the furnace. with an XRS calibrated optical pvrometer or with one that had been compared with it under the conditions of these experiments. With an appropriate window correction, the temperature so measured equals the “black (7) H. W. Goldstein, E. F. Neilson, P. i T.Walsh a n d D. White, J . Phys. Chem., 6 3 , 144R (19.59). (8) M. Hoch and H. L. Johnston, J . A m . Chem. Soc.. 7 6 , 4833 (1954). (9) H. A. J o n e s a n d I. Langmuir, Gen. Elec. Rev., 30, 351 (1927); W. J. W. Edwards, R. Speiser a n d H. L. Johnston. J . A p p l . Phys., 22. 421 (1951). Ta.

August , 1961

VAPORIZATION OF RAREEARTHS

body” temperature of the inside of the crucible within the error of measurement (f3*). Since observation through a side port in one of the furnaces showed that, up to a t least 2200”, the lid was hotter than the main body of the crucible, it can be considered that the measured temperature represents the actual sample temperature. The procedure for a typical Knudsen effusion expriment was as follows. A cell was first degassed a t 2400 for one hour with the pressure in the furnace below 10-6mm. After cooling, the crucible was removed in an argon atmosphere, weighed, and loaded with several hundred milligrams of freshly dried rare earth oxide. Owing to the shrinkage incident on sintering, it was found useful to heat the sample several times at increasing temperature between 1500 and 2000’ and to powder it between heatings. Once the bulk had been reduced to the point where no further shrinkage was observed, the sample was handled only in an argon atmosphere or in VQCUO. The effusion experiments were conducted in a manner previously described. Constancy of temperature was achieved by manual or automatic control of the output of the radiofrequency generator. For both La203and XdzOa a series of runs at different temperatures was made on each sample. The temperature was varied irregularly from run to run until reproducible rates of evaporation were observed a t several temperatures (see e.g., Tables I and 11). An X-ray powder diagram of the residue was taken at the completion of the experiment. A series of runs at constant temperature was also made on each material. In addition, several determinations of the extent of reaction with the crucible were made by heating samples to the point where no further loss of weight (other than the “background” due to the vapor pressure of the crucible material, which was determined separately) was observed. In an effort to estab!ish approximately the magnitude of the “vaporization coefficient” for the reactions involved in these vaporizations, runs were made in which the lid was not placed on the crucible, so that the orifice area to surface area ratio was approximately unity. Since blackbody conditions for temperature measurement were not achieved in this arrangement, the temperatures for such measurements were established by comparing the temperature read in a well ( 1 / 3 1 1 diameter, 3/82” deep) in a !/(“ thick graphite insert placed in the bottom of the crucible m t h the apparent temperature as read on the rim of the crucible, which was subsequently used as a fiduciary. Temperatures mpasured in this way are believed to be good to zt loo.

Results and Discussion Lanthanum and neodymium oxides both react with tantalum Iinudsen cells. The extent of the reaction in the La203 case and the reasons why thermodynamic data cannot be derived from these experiments have been discussed previously.lo Similar behavior is exhibited by Kdz03in tantalum crucibles, and this system is further complicated by extensive formation of a neodymium-tantalumoxygen compound of uncertain stoichiometry.ll The data derived from experiments employing tungsten as the Knudsen cell material can, however, be interpreted in a relatively straightforward manner. The isotherms for both La203and I\;d203 showed an initial high rate of vaporization, a region of constant rate of vaporization, and then a decrease in the rate when approximately 85% of the material had reacted. The high initial rates of evaporation can be correlated with an observed preferential loss of oxygen leading to the formation of a slightly reduced phase which reoxidation studies showed to be LazOs.ss. This composition, the A-form sesquioxide structure with lattice parameters identical to that of the stoichiometric ( I O ) W. W. Goldstein, P. N. Walsh and D. White, J . Phys. Chem., 64, 1087 (1960). (11) P. N. Walsh, H. W. Goldstein and D. White, J . Am. Ceram. Soc.. 48, 229 (1960).

1401

compound, was attained at the point where the rate of evaporation becomes constant. Nd,Os similarly reduced to Nd202.g6. X-Ray examination of the solid phases a t various times during and at the end of a series of runs showed that no further reduction took place and that no new solid phases were formed during vaporization. Hence the vaporization of these oxides occurs stoichiometrically at the compositions La202.9s and Ndz02.sa. The drop in the rate of vaporization after approximately 85% of the sample had vaporized undoubtedly was due to diminution of the surface area of the sample. The species vaporing from the Knudsen cells were initially inferred from the observed vaporization rates and subsequently confirmed by mass spectrometric examination of the vapors, utilizing the equipment and procedures described in Rare Earth Oxides II.12 For both lanthanum and neodymium oxide, three species, identified by mass and isotropic abundance as MO+, M+ and Of (M = La or Nd) were observed to be dependent on the temperature of the crucible. The monoxide ions were always present with greatest intensity. As these had appearance potentials of 5-6 e.v. and since no heavier species were observed, they readily may be ascribed to L a 0 and KdO as progenitors. The metal ions, on the other hand, disappeared when the ionizing electron energy was decreased below 15e.v. ; it was concluded therefore that these species were present primarily as a result of fragmentation of the monoxides. Although an appreciable O+ background was always present in the spectrometer, the following observations verified that oxygen atoms were originating from the effusion cell. (1) The intensity of the O+ peak was markedly decreased when the effusion orifice was moved out of alignment with the collimating slits; with the cell in this position, the rare earth peaks were no longer visible. (2) The shape of the O+ peak suggested that it consisted of two components, one in focus and one out of focus with respect to the component of velocity in the direction of the effusing beam. The component having the same focusing behavior as the rare earth ions disappeared when the crucible alignment was altered as described.l3 The “vaporization coefficient,” a,for the over-all vaporization process of each sesquioxide was established by extrapolation to zero orifice area of the rates of weight loss per unit area (at a given temperature) measured with and without the lid in place on the crucible (orifice area/surface area ratios of 0.058 and 1.00, respectively), according to the procedure described by Speiser and Spretnakl* based on the work of Motzfeldt.16 Since the two sets of measurements were necessarily carried out in different temperature ranges, they were converted to a common temperature by extrapolation of the more precise covered cell measurements to the lower temperatures of the open cell measurements. The “vaporization coefficients” so determined were not less than 0.5 for either material (12) D. White, P. N. Walsh, H. W. Goldstein and D. F. Dever, J . Phgs. Chem., 66, 1404 (1961). (13) This foousing characteristic of the Bendix mass spectrometer has been discussed b y W. C. Wiley, Science, 124, 817 (1956). (14) R. Speiser and J. W. Spretnak, in ‘‘Vacuum Metallurgy,” A m . Electrochem. Soc., 1958. (15) K. Motzfeldt. J . Phus. Chem., 69, 139 (1955).

11. U*.GOLDSTEIN, P. N. WALSHAND D. WHITE

1402

(for La203, a, value of unity was obtained). The value unity has been used in all the calculations below. The total weight loss measurements indicated that, for each sesquioxide, 0.1 mole of tungsten per mole of rare earth sesquioxide was vaporized. As no tungsten containing species was observed in the mass spectrometer, it is concluded that this loss occurs by diffusion of oxygen through the tungsten and subsequent vaporization of a tungsten oxide from the exterior surface. A more pronounced example of this behavior, in the case of tantalum crucibles, recently has been described by us.1o In the present case, however, this effect is small enough to be neglected. Its effect on the calculated heats of formation is less than the uncertainty in the experimental rates of effusion. Furthermore, it is not clear whether the tungsten loss occurs uniformly with time. Some observations made in connection with the vaporization of praseodymium oxide16 indicate that a large percentage of the tungsten may be lost near the beginning of the reaction, before the flat in the isotherm is reached. Since for thermodynamic calculations the only suitable results are those derived under conditions where the vaporization is invariant a t constant temperature, the tungsten loss may not be an important factor at all in the reaction. The observed rates of evaporation a t various temperatures (in the range where the vaporization is invariant a t constant temperature) obtained using Knudsen cells having an orifice area/surface area ratio of 0.058 are shown in Table I for lanthanum oxide and in Table I1 for neodymium oxide. From the experiments discussed above, it is clear that the vaporization for both Laz03 and Nd203 proceeds for all practical purposes as MZOa(c) --ic

2MO(g)

+ O(g)

pvl = La,

Nd (1)

(The difference between the constant-boiling and stoichiometric compositions of the sesquioxides has been neglected.) In Tables I and 11, the calculated partial pressures, equilibrium constants, and heats of reaction for the vaporization of La203and Sdz03, respectively, based on reaction 1 are also given. The partial pressures were calculated from the observed rates of weight loss by the equation14 P,

=

W,rn*(2~RT/M,)l/2

(2)

where Pi, Wi and .Mi are the partial pressure, weight fraction in the vapor and molecular weight, respectively, of the ith vapor species, m* is the observed rate of weight loss corrected by the method of Motzfeldt, T is the absolute temperature, and R is the gas content. The weight fractions were chosen so as to maintain an oxygen/metal ratio of 3//z in the vapor, as required by equation 1. The free energy functions for LazO8were derived from heat capacity7V1'measurements. Free energy functions for 0 were taken from Stull and Sinkel*; for Lao, they were computed from the spectroscopic data given by Herzberg3 with the rotational con(16) Unpuhlishe-1 results of this Laboratory.

(17) J. 0. Blorneke and W. T. Ziegler, .J. Am. Chem. SOL,75, 5099 (1951)

(18) D. R Strill a n d B. C. Sinke, "Thermodynamic Properties of the ElemPnts " American Chemiral Society, Washington, D. C . , 1956.

Vol. 65

stant assumed equal to that of Ba0.3 The same rotational and vibrational contributions to the free energy functions were used for NdO as for Lao. We have shown in the third paper of the serieslg that the entropy of NdO(g) a t 1900OK. is greater by 3 cal. mole-' deg.-l than that of LaO(g). On the assumption that this difference arises entirely from differences in ground state multiplicities, the same quantity has been added to the electronic free energy function of La0 to yield those of NdO. (If the entropy difference arises from lowlying levels of KdO, which is as probable, then the contribution to the free energy function has been overestimated by an unknown amount. In the absence of spectroscopic data on NdO, the question cannot be settled a t present.) The free energy functions for Nd203 reported by us previously7 have been recalculated to take into account some recent measurements by WestrumZ0in the range 4-16'K., which show 8 1 6 = 1.651 cal. mole-' deg.-l and HI6O - Ho0 = 15.74 cal. mole-l. As a result, the free energy functions of Nd203were taken as 56.86, 65.79 and 73.19 cal. mole-' deg.-* a t 1500,2000 and 2500°K., respectively. The heats of formation of LaO(g) and KdO(g) at OOK. were calculated from the heats of reaction shown in Tables I and 11, using the known dissociation energy of 0221 and the heats of formation of lanthanum and neodymium oxides reported by Huber and HolleyZ2corrected to OOK. with the aid of heat content function for the oxides7 and elements.lS The recent, quite different, heat of formation reported by von WartenbergZ3for Laz03 actually agrees well with the value chosen if the most recent heat of solution of La24is used to recalculate theresults. The heats of formation (aHoO) so derived are: Lao, -30.7 kcal. mole-l; NdO, -30.3 kcal. mole-'. The uncertainty in these values arising from experimental factors is indicated by the scatter of the data in Tables I and 11; that arising from the use of approximate free energy functions is estimated to be =k3 kcal. mole-' for La0 and =k 5 kcal. mole-' for NdO. An independent check of these values for the heats of formation was obtained from mass spectrometric investigation of the temperature dependence of the MO+ peaks in vaporization of the solid sesquioxides. Details of the procedure are given in a following article.12 The results are shown in Fig. 1, for lanthanum and neodymium oxides, respectively. In this figure log IMO+Tis plotted against the reciprocal of the absolute temperature. For these measurements an ionizing energy of 25 e.v. was employed. To calculate the heat of reaction, it is assumed that the partial pressure of oxygen is directly proportional to that of the monoxide over the entire temperature range. This is essentially the same approximation used in analyzing the effusion data. Thus, a t the mean tempera(19) P. N. Walsh, D. F. Dever and D. White, J. Phys. Chem., 65, 1410 (1961). (20) E. F. Westrum. private communication. (21) P. Brix a n d B. Herzberg, Can. J. Phys., 5 2 , 110 (1954). (22) E. J. Huber, Jr., and C. E. Holley, Jr., J . -4m. Chem. Soc., 74, 5530 (1952), NdrOa; 7 6 , 3594 (1953), LazOs. (23) H. Von Wartenberg, 2. anorg. allgem. Chem., 299, 227 (1959). (24) F. H. Spedding and J. P. Flynn, J . Am. Chem. Soc., 7 6 , 1474

(1954).

August, 1961

VAPORIZATION O F

RAREEA4RTHS

TABLE I CALCULATION OF HEATOF REACTION LalOs(c) -+ 2La0( g) Rate of effusion, g./cm.a/Bec.

Temp., OK.

2234 2307 2353 2412 2441 2372 2372 2372 2372 2372 2321 2372 2372 2397 2321 2372" a

Po,

PLaO,

x 104 0.595 1.266 2.814 4.482 6.517 3.499 3.414 2.819 3.482 3.301 1.691 3.117 3.029 4.131 1.691 3.117

atm. X 100

atm. X

- R In K.,

4.844 10.48 23.51 37.92 55.46 29.35 28.64 23.65 29.21 27.69 14.04 26.15 25.41 34.84 14.04 26.15

0.7768 1.680 3.770 6.081 8.895 4.707 4.593 3.793 4.685 4.441 2.251 4.193 4.075 5.587 2.251 4.193

76.607 71.996 67.188 64.338 62.070 65.864 66.011 67.154 65.895 66.213 70.263 66.555 66.725 64.842 70.263 66.555

106

Average of seven runs agreeing within & 10%. I

'

l

'

l

'

l

'

l

'

l

-

-

-

0.40-

0.20 -

-

0.00 -

-

-

-0.20-

-

'

- A(Fo--HoO)/R, cal. mole-1

AHoD,

deg. -1

kcal. mole - 1

113.01 112.75 112.58 112.37 112.27 112.52 112.52 112.52 112.52 112.52 112.70 112.52 112.52 112.43 112.70 112.52

423.61 426,22 423.00 426.22 425.56 423.12 423.47 426.17 423.21 423.97 424.67 424.75 425.18 424.92 424.67 424.75

-

kcal. mole-', and AHoo= 216.3 f 6.2 kcal. mole-' for M = La, Kd, respectively. The corresponding heats of formation (AHoO)of LaO(g) and r\TdO(g) are -25.5 f 7 and -28.9 i 7 kcal. mole-', respectively.25 The agreement with the values derived from effusion studies is seen to be satisfactory. The heats of formation obtained by averaging the effusion and mass spectrometric results, each weighted in proportion according to the experimental precision are: LaO(g), AHoo(f) = 29.8 f 4 kcal. mole-'; NdO(g), AHoo(f)= -30.0 f 6 kcal. mole-'. The heat of formation of LaO(g).determined here may be compared with that derived by Chupka, et aE.,2from a mass spectrometric study of the reaction l/ZLa(g)

+ '/nLazOdc) --+ LaO(g)

(4)

For the purpose of this comparison, the free energies of reaction 4 reported by these investigations have been used in conjunction with the free energy functions of La20,(c) and LaO(g) described previously and those of La(g) listed by Stull and Sinke.'* This yields AH$ (reaction 4) = 83.6 f 9 kcal. mole-'. With the heat of sublimation of La given by Daane and SpeddingZ6corrected to absolute zero18 (yielding AHoo(subl) = 97.4 f 1.0 kcal. mole-'), this gives AHoo(f) = -26.2 f 9 kcal. mole-l for LaO(g). This is in agreement with the value derived from the study reported here. The dissociation energies (Doo)of the monoxides can be computed from the heat's of formation, using the heats of sublimation (AHoo)of the rare earth elements and the dissociation energy of oxygen.21 For lanthanum, the heat of sublimation previously

i,o

x IO~("K-'),

Fig. 1.-Variation of pressure of L a 0 ( I L o T ) over LazOa(c) and KdO ( I h o T )over NdzOdc) as a function of reciprocal temperature.

ture, the 'lope times 2'303R times 3/2 is to the heat Of the reaction. The heats of reaction 3 (M = L&,Nd) (3) I/&~O,(~) MO(@;)+ a t the average temperatures are: M = La, A H 2 1 4 0 = 209.2 f 4.2 kcal. mole-1; M = Nd, AHzso,= 204.5 f 4.4 kcal. correcting to the abzero, using heat 'Ontent functions from the sources cited above, yields AHo0 = 217.6 f 6.0 ----f

+ O(g)

Av. 424.65 f 0.98

-

+

1403

(25) T h e heats of reaction 3 calculated from t h e mass spectroscopic d a t a are higher t h a n those calculated from the effnsion results. This m a y be another indication10 of t h e fact t h a t the equilibrium oxygen pressure in these reactions is lower t h a n t h a t ealrulated assuming stoichiometric decomposition of the solids. It was, however, impossible t o determine t h e ratio, MO/O, in t h e vapor phase ma89 spectrometrically with a n y precision due t o t h e oxygen background a n d t h e uncertainty of the relative cross sections of M O + a n d 0'. (26) A. H. Daane and F. H. Spedding, Abstracts, A.C.S. Meeting. Chicago, September, 1958.

D. WHITE,1'. S. WALSH,H. W. GOLDSTEIN A X D D. I?. DEVEII

1404

('.II,CC.lATION OF IIEAT O F

Temp., OK.

2255 228 1 2306 2408 2332 2357 2383 2332 2255 2255 2255 2434 2383 2363 2383 2332 2281 2434 2408"

Hate of effusion. jg./cm.l/sec. x 10'

0.615 0.822 1.040 2.760 1.119 1.521 1.808 1.346 0.561 0.703 0.567 3.089 2.017 1.893 2.060 1.383 0.751 3.156 2.680

TABLE I1 REACTION Nd&(C)

atm. X 10'

PO, atm. X 10'

4.951 6.655 8.467 22.96 9.165 12.53 14.97 11.02 4.519 5.660 4.564 25.85 16.70 15.60 17.05 11.32 6.081 26.41 22.30

0.7900 1.062 1.351 3.665 1.463 1.999 2.388 1.759 0.7211 0.9031 9.7282 4.124 2.665 2.490 2.721 1.807 0.9704 4.214 3.559

PNdO,

+2NdO(g) + o(g) - cal. A(FD-HoD)/T, mole-1

-R In Km

Average of twelve runs agreeing within &IO%.

Vol. G5

76.479 74.716 73.280 67.330 72.808 70.945 69.884 71.708 77.023 76.447 76.964 66.626 69.220 69.639 69.105 71.546 75.253 66.498 67.512

AH@,

deg. -1

kcal. mole-'

112.59 112.46 112.34 111.83 112.21 112.08 111.95 112.21 112.59 112.59 112.59 111.70 111.95 112.05 111.95 115.21 112.46 111.70 111.83

426.35 426.95 428.04 431.42 431.46 431.39 433.31 428.80 427.58 42G. 28 427.44 434.04 431.73 429.33 431.46 428.52 428.18 433.73 431.86

Av. 429.89 f 2.30

given has been used. For neodymium, the value 1.0 kcal. mole-', derived from the heat of 76.6 vaporization determined in this Laboratory12 and heat content functions for the gas18 and condensed phases,*' has been used. The resultant dissociation energies are: Lao, 8.08 f 0.2 e.v., NdO, 7.18 i 0.3 e.v. It is notewort:hy that the heat's of formation of the gaseous monoxides are very nearly the same and that the dissociation energies differ by just the difference in heats of sub1imat)ionof the metal; i.e., the energy released in adding an oxygen atom t80 a free lanthanum atom is greater by just the

*

(37) F. H. Spedding, J. J. McICeown and A. H. D a m e , Chem., 64, 289 (19GO).

J. Phys.

amount that the binding energy of lanthanum metal exceeds that of neodymium. This suggests a correlation between the binding mechanisms in the two cases and indicates a possibility that further study of the monoxides may be useful in explaining the behavior of the rare earth metals. Rather than attempt to develop this further here, we consider it prudent to wait until we have presented evidence relative to additional rare earth syst.ems. Achowledgments.-The authors wish to acknowledge the assistance of Dr. M. Hoch in initially developing this program, and Mr, William E. Housden in carrying out the effusion measurements. H.W.G. wishes to thank the Monsanto Company for a fellowship during One year in which this work was carried on.

RARE I3ARI"S. 11. A MASS SPECTROMETRIC DETERMINATIOX OF THE HEATS OF SUBLIMATION (OR VAPORIZATION) OF NEODYRIIUN, PRASEODYMIURI, GADOLINIUM, TERBIUM, DYSPROSIUM, HOLMIUM, ERBIUM AND LUTETIUM1 BY D.4VID WHITE,

PATRICK

hT. WALSH, HAROLD lv. GOLDSTEIN AND DAVIDIi'. DEVER

Cryogenic Laboratory, Department of Chemistry, The Ohio State University, Columbus, Ohio Recrzved Match 18, 1061

A Time of Flight mass spectrometer has been adapted for thermodynamic investigations at elevated temperatures. The apparatus is described in detail. The heats of sublimation (or vaporization) of several rare earth metals have been determined by the mass spectrometric method, from the variation with temperature of the intensity of an atomic beam effusing from a Knudsen cell, in the temperature range 1253 to 2044'K.

Introduction I n an earlier article,2 the thermodynamics of .\.aporization of the sesquioxides LanOs(c) and (1) This

work was supported by the Office of Naval Research. (2) H. W. Goldstein, P. N. Walsh and David White, J . Phys.-Chem., 65. 1400 (1961).

NdlOJc) a t elevated temperatures is discussed. I n order to interpret the Knudsen effusion data and calculate the dissociation energies of the rare earth gaseous monoxides the heats Of of the IMh& must be known. For nearly all the sesquioxides that have been studied to date, it has